Prediction Model For Risk Of Breast Cancer Considering Interaction Between The Risk Factors

This paper focuses on expansion of Barlows predictive model in which a large data set of 2392998 eligible screening mammograms taken from Breast Cancer Surveillance Consortium which was previously used by Barlow in 2006 to predict a diagnosis of breast cancer in women through including interaction of exploratory variables. 12 explanatory variables that are assumed to influence the risk of developing breast cancer in women and they are age breast density menopause status race Hispanic BMI number of first degree relatives with breast cancer previous breast procedure age at first birth surgical menopause results of last mammogram and current hormone therapy. Forward selection method was used to select the best predictive model including significant interaction terms. The results showed 33 interactions were included in the new model through forward selection procedure improved the predictive model. However only 10 interaction terms were found to be significant across all levels of the risk factors. Also the updated predictive model was found to better than the main effect model as the AIC value decreased.